[TIKA-3427] Duplicate characters in some words

• Type: Bug
• Status: Closed
• Priority: Minor
• Resolution: Not A Problem
• Affects Version/s: 1.26
• Component/s: tika-server

Environment

Windows 10 x64

• Flags: Important
• Assignee: Unassigned
• Reporter: Sal
• Created: 31/May/21 19:28
• Updated: 01/Jun/21 13:57
• Resolved: 01/Jun/21 13:57

Description

When processing PDF document to extract text using Tika Server, the output contains words with some duplicated characters and partial words.  

I am sending the PDF using a POST request to the Tika Server running locally at url http://localhost:9998/tika with the PDF attached to the body of the message and headers 

Content-Type : application/pdf

X-Tika-PDFextractInlineImages : true

X-Tika-PDFOcrStrategy: ocr_and_text_extraction

An attached PDF document  is provided as an example

The output looks like this, incorrect text is in red text

 

PPAATIENTTIENT

DISEASE Lung cancer (NOS)
NAME
DATE OF BIRTH
SEX Male
MEDICAL RECORD # Not given

PHYPHYSICIANSICIAN

ORDERING PHYSICIAN
MEDICAL FACILITY
ADDITIONAL RECIPIENT None
MEDICAL FACILITY ID
PATHOLOGIST Not Provided

SPESPECIMENCIMEN

SPECIMEN ID
SPECIMEN TYPE Blood
DATE OF COLLECTION
SPECIMEN RECEIVED
MEDIAN EXON COVERAGE

Biomarker Findings
MSI SMSI Statatus Undettus Undetermined.ermined.